ABSTRACT Recombinant adeno-associated viruses (AAV) are promising vectors for human gene therapy. Despite regulatory approval and a rapidly growing pipeline, clinical data have indicated that vector dose-related hepatotoxicity is a significant concern. Recent preclinical studies suggest that heterogeneity in AAV vector preparations, including the presence of defective interfering particles with partial or truncated genomes contribute to such undesirable side effects. Although somewhat resolvable with anti-inflammatory drug regimens, permanent loss of gene expression has been reported. We postulate that improvements to AAV vector design and composition are likely to improve the safety profile. To achieve the latter goals, a better understanding of particle heterogeneity in AAV vectors packaging different genomes as well as the impact of such on AAV biology is essential. In the current proposal, we will first advance a potentially disruptive technology, charge detection mass spectrometry (CDMS) to help map the mass landscape of vector compositions associated with genome packaging. We will then bridge this basic information with new directions in AAV capsid engineering with the goal of improving compositional homogeneity and reducing the potential for side effects. Each aim is tailored to directly interrogate the proposed correlation between the composition of recombinant vectors with viral infectivity in cell culture as well as gene transfer efficiency and toxicity in mouse models. We expect our efforts to (i) offer significant, new insight into parvoviral capsid-genome interactions, (ii) enable the development and validation of a novel analytical technique for rapid screening and quality testing of clinical AAV samples, and (iii) improve AAV gene transfer efficiency and safety profile through capsid engineering.